vrobbler/vrobbler/apps/scrobbles/models.py

import pytz
import calendar
import datetime
import logging
from decimal import Decimal
from typing import Iterable, Optional
from uuid import uuid4

import pendulum
from boardgames.models import BoardGame
from books.koreader import process_koreader_sqlite_file
from books.models import Book
from django.conf import settings
from django.contrib.auth import get_user_model
from django.db import models
from django.urls import reverse
from django.utils import timezone
from django.utils.functional import cached_property
from django_extensions.db.models import TimeStampedModel
from imagekit.models import ImageSpecField
from imagekit.processors import ResizeToFit
from locations.models import GeoLocation
from music.lastfm import LastFM
from music.models import Artist, Track
from podcasts.models import PodcastEpisode
from profiles.utils import (
    end_of_day,
    end_of_month,
    end_of_week,
    start_of_day,
    start_of_month,
    start_of_week,
)
from scrobbles.constants import LONG_PLAY_MEDIA
from scrobbles.stats import build_charts
from scrobbles.utils import media_class_to_foreign_key
from sports.models import SportEvent
from videogames import retroarch
from videogames.models import VideoGame
from videos.models import Series, Video
from scrobbles.dataclasses import (
    BoardGameMetadata,
    LifeEventMetadata,
    ScrobbleMetadataDecoder,
    ScrobbleMetadataEncoder,
    VideoMetadata,
)
from webpages.models import WebPage
from lifeevents.models import LifeEvent

from vrobbler.apps.scrobbles.constants import MEDIA_END_PADDING_SECONDS

logger = logging.getLogger(__name__)
User = get_user_model()
BNULL = {"blank": True, "null": True}

POINTS_FOR_MOVEMENT_HISTORY = int(
    getattr(settings, "POINTS_FOR_MOVEMENT_HISTORY", 3)
)


class BaseFileImportMixin(TimeStampedModel):
    user = models.ForeignKey(User, on_delete=models.DO_NOTHING, **BNULL)
    uuid = models.UUIDField(editable=False, default=uuid4)
    processing_started = models.DateTimeField(**BNULL)
    processed_finished = models.DateTimeField(**BNULL)
    process_log = models.TextField(**BNULL)
    process_count = models.IntegerField(**BNULL)

    class Meta:
        abstract = True

    def __str__(self):
        return f"{self.import_type} import on {self.human_start}"

    @property
    def human_start(self):
        start = "Unknown"
        if self.processing_started:
            start = self.processing_started.strftime("%B %d, %Y at %H:%M")
        return start

    @property
    def import_type(self) -> str:
        return "Unknown Import Source"

    def process(self, force=False):
        logger.warning("Process not implemented")

    def undo(self, dryrun=False):
        """Accepts the log from a scrobble import and removes the scrobbles"""
        from scrobbles.models import Scrobble

        if not self.process_log:

            logger.warning("No lines in process log found to undo")
            return

        for line in self.process_log.split("\n"):
            scrobble_id = line.split("\t")[0]
            scrobble = Scrobble.objects.filter(id=scrobble_id).first()
            if not scrobble:
                logger.warning(
                    f"Could not find scrobble {scrobble_id} to undo"
                )
                continue
            logger.info(f"Removing scrobble {scrobble_id}")
            if not dryrun:
                scrobble.delete()
        self.processed_finished = None
        self.processing_started = None
        self.process_count = None
        self.process_log = ""
        self.save(
            update_fields=[
                "processed_finished",
                "processing_started",
                "process_log",
                "process_count",
            ]
        )

    def scrobbles(self) -> models.QuerySet:
        scrobble_ids = []
        if self.process_log:
            for line in self.process_log.split("\n"):
                sid = line.split("\t")[0]
                if sid:
                    scrobble_ids.append(sid)
        return Scrobble.objects.filter(id__in=scrobble_ids)

    def mark_started(self):
        self.processing_started = timezone.now()
        self.save(update_fields=["processing_started"])

    def mark_finished(self):
        self.processed_finished = timezone.now()
        self.save(update_fields=["processed_finished"])

    def record_log(self, scrobbles):
        self.process_log = ""
        if not scrobbles:
            self.process_count = 0
            self.save(update_fields=["process_log", "process_count"])
            return

        for count, scrobble in enumerate(scrobbles):
            scrobble_str = f"{scrobble.id}\t{scrobble.timestamp}\t{scrobble.media_obj.title}"
            log_line = f"{scrobble_str}"
            if count > 0:
                log_line = "\n" + log_line
            self.process_log += log_line
        self.process_count = len(scrobbles)
        self.save(update_fields=["process_log", "process_count"])

    @property
    def upload_file_path(self):
        raise NotImplementedError


class KoReaderImport(BaseFileImportMixin):
    class Meta:
        verbose_name = "KOReader Import"

    @property
    def import_type(self) -> str:
        return "KOReader"

    def get_absolute_url(self):
        return reverse(
            "scrobbles:koreader-import-detail", kwargs={"slug": self.uuid}
        )

    def get_path(instance, filename):
        extension = filename.split(".")[-1]
        uuid = instance.uuid
        return f"koreader-uploads/{uuid}.{extension}"

    @property
    def upload_file_path(self) -> str:
        if getattr(settings, "USE_S3_STORAGE"):
            path = self.sqlite_file.url
        else:
            path = self.sqlite_file.path
        return path

    sqlite_file = models.FileField(upload_to=get_path, **BNULL)

    def process(self, force=False):

        if self.processed_finished and not force:
            logger.info(
                f"{self} already processed on {self.processed_finished}"
            )
            return

        self.mark_started()
        scrobbles = process_koreader_sqlite_file(
            self.upload_file_path, self.user.id
        )
        self.record_log(scrobbles)
        self.mark_finished()


class AudioScrobblerTSVImport(BaseFileImportMixin):
    class Meta:
        verbose_name = "AudioScrobbler TSV Import"

    @property
    def import_type(self) -> str:
        return "AudiosScrobbler"

    def get_absolute_url(self):
        return reverse(
            "scrobbles:tsv-import-detail", kwargs={"slug": self.uuid}
        )

    def get_path(instance, filename):
        extension = filename.split(".")[-1]
        uuid = instance.uuid
        return f"audioscrobbler-uploads/{uuid}.{extension}"

    @property
    def upload_file_path(self):
        if getattr(settings, "USE_S3_STORAGE"):
            path = self.tsv_file.url
        else:
            path = self.tsv_file.path
        return path

    tsv_file = models.FileField(upload_to=get_path, **BNULL)

    def process(self, force=False):
        from scrobbles.tsv import process_audioscrobbler_tsv_file

        if self.processed_finished and not force:
            logger.info(
                f"{self} already processed on {self.processed_finished}"
            )
            return

        self.mark_started()

        tz = None
        user_id = None
        if self.user:
            user_id = self.user.id
            tz = self.user.profile.tzinfo
        scrobbles = process_audioscrobbler_tsv_file(
            self.upload_file_path, user_id, user_tz=tz
        )
        self.record_log(scrobbles)
        self.mark_finished()


class LastFmImport(BaseFileImportMixin):
    class Meta:
        verbose_name = "Last.FM Import"

    @property
    def import_type(self) -> str:
        return "LastFM"

    def get_absolute_url(self):
        return reverse(
            "scrobbles:lastfm-import-detail", kwargs={"slug": self.uuid}
        )

    def process(self, import_all=False):
        """Import scrobbles found on LastFM"""
        if self.processed_finished:
            logger.info(
                f"{self} already processed on {self.processed_finished}"
            )
            return

        last_import = None
        if not import_all:
            try:
                last_import = LastFmImport.objects.exclude(id=self.id).last()
            except:
                pass

        if not import_all and not last_import:
            logger.warn(
                "No previous import, to import all Last.fm scrobbles, pass import_all=True"
            )
            return

        lastfm = LastFM(self.user)
        last_processed = None
        if last_import:
            last_processed = last_import.processed_finished

        self.mark_started()

        scrobbles = lastfm.import_from_lastfm(last_processed)

        self.record_log(scrobbles)
        self.mark_finished()


class RetroarchImport(BaseFileImportMixin):
    class Meta:
        verbose_name = "Retroarch Import"

    @property
    def import_type(self) -> str:
        return "Retroarch"

    def get_absolute_url(self):
        return reverse(
            "scrobbles:retroarch-import-detail", kwargs={"slug": self.uuid}
        )

    def process(self, import_all=False, force=False):
        """Import scrobbles found on Retroarch"""
        if self.processed_finished and not force:
            logger.info(
                f"{self} already processed on {self.processed_finished}"
            )
            return

        if force:
            logger.info(f"You told me to force import from Retroarch")

        if not self.user.profile.retroarch_path:
            logger.info(
                "Tying to import Retroarch logs, but user has no retroarch_path configured"
            )

        self.mark_started()

        scrobbles = retroarch.import_retroarch_lrtl_files(
            self.user.profile.retroarch_path,
            self.user.id,
        )

        self.record_log(scrobbles)
        self.mark_finished()


class ChartRecord(TimeStampedModel):
    """Sort of like a materialized view for what we could dynamically generate,
    but would kill the DB as it gets larger. Collects time-based records
    generated by a cron-like archival job

    1972 by Josh Rouse - #3 in 2023, January

    """

    user = models.ForeignKey(User, on_delete=models.DO_NOTHING, **BNULL)
    rank = models.IntegerField(db_index=True)
    count = models.IntegerField(default=0)
    year = models.IntegerField(**BNULL)
    month = models.IntegerField(**BNULL)
    week = models.IntegerField(**BNULL)
    day = models.IntegerField(**BNULL)
    video = models.ForeignKey(Video, on_delete=models.DO_NOTHING, **BNULL)
    series = models.ForeignKey(Series, on_delete=models.DO_NOTHING, **BNULL)
    artist = models.ForeignKey(Artist, on_delete=models.DO_NOTHING, **BNULL)
    track = models.ForeignKey(Track, on_delete=models.DO_NOTHING, **BNULL)
    period_start = models.DateTimeField(**BNULL)
    period_end = models.DateTimeField(**BNULL)

    def save(self, *args, **kwargs):
        profile = self.user.profile

        if self.week:
            # set start and end to start and end of week
            period = datetime.date.fromisocalendar(self.year, self.week, 1)
            self.period_start = start_of_week(period, profile)
            self.period_start = end_of_week(period, profile)
        if self.day:
            period = datetime.datetime(self.year, self.month, self.day)
            self.period_start = start_of_day(period, profile)
            self.period_end = end_of_day(period, profile)
        if self.month and not self.day:
            period = datetime.datetime(self.year, self.month, 1)
            self.period_start = start_of_month(period, profile)
            self.period_end = end_of_month(period, profile)
        super(ChartRecord, self).save(*args, **kwargs)

    @property
    def media_obj(self):
        media_obj = None
        if self.video:
            media_obj = self.video
        if self.track:
            media_obj = self.track
        if self.artist:
            media_obj = self.artist
        return media_obj

    @property
    def month_str(self) -> str:
        month_str = ""
        if self.month:
            month_str = calendar.month_name[self.month]
        return month_str

    @property
    def day_str(self) -> str:
        day_str = ""
        if self.day:
            day_str = str(self.day)
        return day_str

    @property
    def week_str(self) -> str:
        week_str = ""
        if self.week:
            week_str = str(self.week)
        return "Week " + week_str

    @property
    def period(self) -> str:
        period = str(self.year)
        if self.month:
            period = " ".join([self.month_str, period])
        if self.week:
            period = " ".join([self.week_str, period])
        if self.day:
            period = " ".join([self.day_str, period])
        return period

    @property
    def period_type(self) -> str:
        period = "year"
        if self.month:
            period = "month"
        if self.week:
            period = "week"
        if self.day:
            period = "day"
        return period

    def __str__(self):
        title = f"#{self.rank} in {self.period}"
        if self.day or self.week:
            title = f"#{self.rank} on {self.period}"
        return title

    def link(self):
        get_params = f"?date={self.year}"
        if self.week:
            get_params = get_params = get_params + f"-W{self.week}"
        if self.month:
            get_params = get_params = get_params + f"-{self.month}"
        if self.day:
            get_params = get_params = get_params + f"-{self.day}"
        if self.artist:
            get_params = get_params + "&media=Artist"
        return reverse("scrobbles:charts-home") + get_params

    @classmethod
    def build(cls, user, **kwargs):
        build_charts(user=user, **kwargs)

    @classmethod
    def for_year(cls, user, year):
        return cls.objects.filter(year=year, user=user)

    @classmethod
    def for_month(cls, user, year, month):
        return cls.objects.filter(year=year, month=month, user=user)

    @classmethod
    def for_day(cls, user, year, day, month):
        return cls.objects.filter(year=year, month=month, day=day, user=user)

    @classmethod
    def for_week(cls, user, year, week):
        return cls.objects.filter(year=year, week=week, user=user)


class Scrobble(TimeStampedModel):
    """A scrobble tracks played media items by a user."""

    class MediaType(models.TextChoices):
        """Enum mapping a media model type to a string"""

        VIDEO = "Video", "Video"
        TRACK = "Track", "Track"
        PODCAST_EPISODE = "PodcastEpisode", "Podcast episode"
        SPORT_EVENT = "SportEvent", "Sport event"
        BOOK = "Book", "Book"
        VIDEO_GAME = "VideoGame", "Video game"
        BOARD_GAME = "BoardGame", "Board game"
        GEO_LOCATION = "GeoLocation", "GeoLocation"
        WEBPAGE = "WebPage", "Web Page"
        LIFE_EVENT = "LifeEvent", "Life event"

    uuid = models.UUIDField(editable=False, **BNULL)
    video = models.ForeignKey(Video, on_delete=models.DO_NOTHING, **BNULL)
    track = models.ForeignKey(Track, on_delete=models.DO_NOTHING, **BNULL)
    podcast_episode = models.ForeignKey(
        PodcastEpisode, on_delete=models.DO_NOTHING, **BNULL
    )
    sport_event = models.ForeignKey(
        SportEvent, on_delete=models.DO_NOTHING, **BNULL
    )
    book = models.ForeignKey(Book, on_delete=models.DO_NOTHING, **BNULL)
    video_game = models.ForeignKey(
        VideoGame, on_delete=models.DO_NOTHING, **BNULL
    )
    board_game = models.ForeignKey(
        BoardGame, on_delete=models.DO_NOTHING, **BNULL
    )
    geo_location = models.ForeignKey(
        GeoLocation, on_delete=models.DO_NOTHING, **BNULL
    )
    web_page = models.ForeignKey(WebPage, on_delete=models.DO_NOTHING, **BNULL)
    life_event = models.ForeignKey(
        LifeEvent, on_delete=models.DO_NOTHING, **BNULL
    )
    media_type = models.CharField(
        max_length=14, choices=MediaType.choices, default=MediaType.VIDEO
    )
    user = models.ForeignKey(
        User, blank=True, null=True, on_delete=models.DO_NOTHING
    )

    # Time keeping
    timestamp = models.DateTimeField(**BNULL)
    stop_timestamp = models.DateTimeField(**BNULL)
    playback_position_seconds = models.IntegerField(**BNULL)

    # Status indicators
    is_paused = models.BooleanField(default=False)
    played_to_completion = models.BooleanField(default=False)
    in_progress = models.BooleanField(default=True)

    # Metadata
    source = models.CharField(max_length=255, **BNULL)
    log = models.JSONField(
        **BNULL,
        encoder=ScrobbleMetadataEncoder,
        decoder=ScrobbleMetadataDecoder,
    )
    timezone = models.CharField(max_length=50, **BNULL)

    # Fields for keeping track of book data
    book_koreader_hash = models.CharField(max_length=50, **BNULL)
    book_pages_read = models.IntegerField(**BNULL)
    book_page_data = models.JSONField(**BNULL)

    # Fields for keeping track of video game data
    videogame_save_data = models.FileField(
        upload_to="scrobbles/videogame_save_data/", **BNULL
    )
    videogame_screenshot = models.ImageField(
        upload_to="scrobbles/videogame_screenshot/", **BNULL
    )
    videogame_screenshot_small = ImageSpecField(
        source="videogame_screenshot",
        processors=[ResizeToFit(100, 100)],
        format="JPEG",
        options={"quality": 60},
    )
    videogame_screenshot_medium = ImageSpecField(
        source="videogame_screenshot",
        processors=[ResizeToFit(300, 300)],
        format="JPEG",
        options={"quality": 75},
    )
    long_play_seconds = models.BigIntegerField(**BNULL)
    long_play_complete = models.BooleanField(**BNULL)

    def save(self, *args, **kwargs):
        if not self.uuid:
            self.uuid = uuid4()

        if not self.timezone:
            timezone = settings.TIME_ZONE
            if self.user and self.user.profile:
                timezone = self.user.profile.timezone
            self.timezone = timezone

        # Microseconds mess up Django's filtering, and we don't need be that specific
        if self.timestamp:
            self.timestamp = self.timestamp.replace(microsecond=0)
        self.media_type = self.MediaType(self.media_obj.__class__.__name__)

        return super(Scrobble, self).save(*args, **kwargs)

    def push_to_archivebox(self):
        pushable_media = hasattr(
            self.media_obj, "push_to_archivebox"
        ) and callable(self.media_obj.push_to_archivebox)

        if pushable_media and self.user.profile.archivebox_url:
            try:
                self.media_obj.push_to_archivebox(
                    url=self.user.profile.archivebox_url,
                    username=self.user.profile.archivebox_username,
                    password=self.user.profile.archivebox_password,
                )
            except Exception:
                logger.info(
                    "Failed to push URL to archivebox",
                    extra={
                        "archivebox_url": self.user.profile.archivebox_url,
                        "archivebox_username": self.user.profile.archivebox_username,
                    },
                )

    @property
    def metadata(self):
        metadata_cls = None
        if self.media_type == self.MediaType.LIFE_EVENT:
            metadata_cls = LifeEventMetadata
        if self.media_type == self.MediaType.BOARD_GAME:
            metadata_cls = BoardGameMetadata
        if self.media_type == self.MediaType.VIDEO:
            metadata_cls = VideoMetadata

        if not metadata_cls:
            logger.warn(
                f"Media type has no metadata class",
                extra={"media_type": self.media_type, "scrobble_id": self.id},
            )
            return None

        return metadata_cls.from_dict(self.log)

    @property
    def tzinfo(self):
        return pytz.timezone(self.timezone)

    @property
    def scrobble_media_key(self) -> str:
        return media_class_to_foreign_key(self.media_type) + "_id"

    @property
    def status(self) -> str:
        if self.is_paused:
            return "paused"
        if self.played_to_completion:
            return "finished"
        if self.in_progress:
            return "in-progress"
        return "zombie"

    @property
    def is_stale(self) -> bool:
        """Mark scrobble as stale if it's been more than an hour since it was updated"""
        is_stale = False
        now = timezone.now()
        seconds_since_last_update = (now - self.modified).seconds
        if seconds_since_last_update >= self.media_obj.SECONDS_TO_STALE:
            is_stale = True
        return is_stale

    @property
    def previous(self) -> "Scrobble":
        return (
            self.media_obj.scrobble_set.order_by("-timestamp")
            .filter(timestamp__lt=self.timestamp)
            .first()
        )

    @property
    def next(self) -> "Scrobble":
        return (
            self.media_obj.scrobble_set.order_by("timestamp")
            .filter(timestamp__gt=self.timestamp)
            .first()
        )

    @property
    def previous_by_media(self) -> "Scrobble":
        return (
            Scrobble.objects.filter(
                media_type=self.media_type,
                user=self.user,
                timestamp__lt=self.timestamp,
            )
            .order_by("-timestamp")
            .first()
        )

    @property
    def next_by_media(self) -> "Scrobble":
        return (
            Scrobble.objects.filter(
                media_type=self.media_type,
                user=self.user,
                timestamp__gt=self.timestamp,
            )
            .order_by("-timestamp")
            .first()
        )

    @property
    def previous_by_user(self) -> "Scrobble":
        return (
            Scrobble.objects.order_by("-timestamp")
            .filter(timestamp__lt=self.timestamp)
            .first()
        )

    @property
    def next_by_user(self) -> "Scrobble":
        return (
            Scrobble.objects.order_by("-timestamp")
            .filter(timestamp__gt=self.timestamp)
            .first()
        )

    @property
    def session_pages_read(self) -> Optional[int]:
        if not self.book_pages_read:
            return 0
        return self.book_pages_read

    @property
    def is_long_play(self) -> bool:
        return self.media_obj.__class__.__name__ in LONG_PLAY_MEDIA.values()

    @property
    def percent_played(self) -> int:
        if not self.media_obj:
            return 0

        if self.media_obj and not self.media_obj.run_time_seconds:
            return 100

        if not self.playback_position_seconds and self.played_to_completion:
            return 100

        playback_seconds = self.playback_position_seconds
        if not playback_seconds:
            playback_seconds = (timezone.now() - self.timestamp).seconds

        run_time_secs = self.media_obj.run_time_seconds
        percent = int((playback_seconds / run_time_secs) * 100)

        if self.is_long_play:
            long_play_secs = 0
            if self.previous and not self.previous.long_play_complete:
                long_play_secs = self.previous.long_play_seconds or 0
            percent = int(
                ((playback_seconds + long_play_secs) / run_time_secs) * 100
            )

        return percent

    @property
    def probably_still_in_progress(self) -> bool:
        """Add our start time to our media run time to get when we expect to

        Audio tracks should be given a second or two of grace, videos should
        be given closer to 30 minutes, because the odds of watching it back to
        back are very slim.

        """
        is_in_progress = False
        padding_seconds = MEDIA_END_PADDING_SECONDS.get(self.media_type)
        if not padding_seconds:
            return is_in_progress

        expected_end = self.timestamp + datetime.timedelta(
            seconds=self.media_obj.run_time_seconds
        )
        expected_end_padded = expected_end + datetime.timedelta(
            seconds=padding_seconds
        )
        # Take our start time, add our media length and an extra 30 min (1800s) is it still in the future? keep going
        is_in_progress = expected_end_padded > pendulum.now()
        logger.info(
            "[scrobbling] checking if we're probably still playing",
            extra={
                "media_id": self.media_obj.id,
                "scrobble_id": self.id,
                "media_type": self.media_type,
                "probably_still_in_progress": is_in_progress,
            },
        )
        return is_in_progress

    @property
    def can_be_updated(self) -> bool:
        if self.media_obj.__class__.__name__ in LONG_PLAY_MEDIA.values():
            logger.info(
                "[scrobbling] cannot be updated, long play media",
                extra={
                    "media_id": self.media_obj.id,
                    "scrobble_id": self.id,
                    "media_type": self.media_type,
                },
            )
            return False

        if self.percent_played >= 100 and not self.probably_still_in_progress:
            logger.info(
                "[scrobbling] cannot be updated, existing scrobble is 100% played",
                extra={
                    "media_id": self.media_obj.id,
                    "scrobble_id": self.id,
                    "media_type": self.media_type,
                },
            )
            return False

        if self.is_stale:
            logger.info(
                "[scrobbling] cannot be udpated, stale",
                extra={
                    "media_id": self.media_obj.id,
                    "scrobble_id": self.id,
                    "media_type": self.media_type,
                },
            )
            return False
        logger.info(
            "[scrobbling] can be updated",
            extra={
                "media_id": self.media_obj.id,
                "scrobble_id": self.id,
                "media_type": self.media_type,
            },
        )
        return True

    @property
    def media_obj(self):
        media_obj = None
        if self.video:
            media_obj = self.video
        if self.track:
            media_obj = self.track
        if self.podcast_episode:
            media_obj = self.podcast_episode
        if self.sport_event:
            media_obj = self.sport_event
        if self.book:
            media_obj = self.book
        if self.video_game:
            media_obj = self.video_game
        if self.board_game:
            media_obj = self.board_game
        if self.geo_location:
            media_obj = self.geo_location
        if self.web_page:
            media_obj = self.web_page
        if self.life_event:
            media_obj = self.life_event
        return media_obj

    def __str__(self):
        timestamp = self.timestamp.strftime("%Y-%m-%d")
        return f"Scrobble of {self.media_obj} ({timestamp})"

    def calc_reading_duration(self) -> int:
        duration = 0
        if self.book_page_data:
            for k, v in self.book_page_data.items():
                duration += v.get("duration")
        return duration

    def calc_pages_read(self) -> int:
        pages_read = 0
        if self.book_page_data:
            pages = [int(k) for k in self.book_page_data.keys()]
            pages.sort()
            if len(pages) == 1:
                pages_read = 1
            elif len(pages) >= 2:
                pages_read += pages[-1] - pages[0]
            else:
                pages_read = pages[-1] - pages[0]
        return pages_read

    @property
    def last_page_read(self) -> int:
        last_page = 0
        if self.book_page_data:
            pages = [int(k) for k in self.book_page_data.keys()]
            pages.sort()
            last_page = pages[-1]
        return last_page

    @classmethod
    def create_or_update(
        cls, media, user_id: int, scrobble_data: dict, **kwargs
    ) -> "Scrobble":
        key = media_class_to_foreign_key(media.__class__.__name__)
        media_query = models.Q(**{key: media})
        scrobble_data[key + "_id"] = media.id

        # Find our last scrobble of this media item (track, video, etc)
        scrobble = (
            cls.objects.filter(
                media_query,
                user_id=user_id,
            )
            .order_by("-timestamp")
            .first()
        )
        source = scrobble_data.get("source")
        mtype = media.__class__.__name__
        mopidy_status = scrobble_data.get("mopidy_status", None)

        # GeoLocations are a special case scrobble
        if mtype == cls.MediaType.GEO_LOCATION:
            logger.warn(
                f"[scrobbling] use create_or_update_location for GeoLocations"
            )
            scrobble = cls.create_or_update_location(
                media, scrobble_data, user_id
            )
            return scrobble

        logger.info(
            f"[scrobbling] check for existing scrobble to update ",
            extra={
                "scrobble_id": scrobble.id if scrobble else None,
                "media_type": mtype,
                "media_id": media.id,
                "scrobble_data": scrobble_data,
                "percent_played": scrobble.percent_played if scrobble else 0,
                "can_be_updated": scrobble.can_be_updated
                if scrobble
                else False,
            },
        )

        scrobble_data["playback_status"] = scrobble_data.pop(
            "mopidy_status", scrobble_data.pop("jellyfin_status", None)
        )
        # If it's marked as stopped, send it through our update mechanism, which will complete it
        if scrobble and (
            scrobble.can_be_updated
            or scrobble_data["playback_status"] == "stopped"
        ):
            return scrobble.update(scrobble_data)

        # Discard status before creating
        scrobble_data.pop("playback_status")
        logger.info(
            f"[scrobbling] creating new scrobble",
            extra={
                "scrobble_id": scrobble.id if scrobble else None,
                "media_type": mtype,
                "media_id": media.id,
                "source": source,
            },
        )

        return cls.create(scrobble_data)

    @classmethod
    def create_or_update_location(
        cls, location: GeoLocation, scrobble_data: dict, user_id: int
    ) -> "Scrobble":
        """Location is special type, where the current scrobble for a user is always the
        current active scrobble, and we only finish it a move on if we get a new location
        that is far enough (and far enough over the last three past scrobbles) to have
        actually moved.
        """
        key = media_class_to_foreign_key(location.__class__.__name__)
        scrobble_data[key + "_id"] = location.id

        scrobble = (
            cls.objects.filter(
                media_type=cls.MediaType.GEO_LOCATION,
                user_id=user_id,
                timestamp__lte=scrobble_data.get("timestamp"),
            )
            .order_by("-timestamp")
            .first()
        )

        logger.info(
            f"[scrobbling] fetching last location scrobble",
            extra={
                "scrobble_id": scrobble.id if scrobble else None,
                "media_type": cls.MediaType.GEO_LOCATION,
                "media_id": location.id,
                "scrobble_data": scrobble_data,
            },
        )

        if not scrobble:
            logger.info(
                f"[scrobbling] finished - no existing location scrobbles found",
                extra={
                    "media_id": location.id,
                    "media_type": cls.MediaType.GEO_LOCATION,
                },
            )
            return cls.create(scrobble_data)

        if scrobble.media_obj == location:
            logger.info(
                f"[scrobbling] finished - same location - not moved",
                extra={
                    "media_type": cls.MediaType.GEO_LOCATION,
                    "media_id": location.id,
                    "scrobble_id": scrobble.id,
                    "scrobble_media_id": scrobble.media_obj.id,
                },
            )
            return scrobble

        has_moved = location.has_moved(scrobble.media_obj)
        logger.info(
            f"[scrobbling] checking - has last location has moved?",
            extra={
                "scrobble_id": scrobble.id,
                "scrobble_media_id": scrobble.media_obj.id,
                "media_type": cls.MediaType.GEO_LOCATION,
                "media_id": location.id,
                "has_moved": has_moved,
            },
        )
        if not has_moved:
            logger.info(
                f"[scrobbling] finished - not from old location - not moved",
                extra={
                    "scrobble_id": scrobble.id,
                    "media_id": location.id,
                    "media_type": cls.MediaType.GEO_LOCATION,
                    "old_media__id": scrobble.media_obj.id,
                },
            )
            return scrobble

        if existing_locations := location.in_proximity(named=True):
            existing_location = existing_locations.first()
            scrobble.log[
                pendulum.now().timestamp
            ] = f"Location {location.id} too close to this scrobble"
            scrobble.save(update_fields=["log"])
            logger.info(
                f"[scrobbling] finished - found existing named location",
                extra={
                    "media_id": location.id,
                    "media_type": cls.MediaType.GEO_LOCATION,
                    "old_media_id": existing_location.id,
                },
            )
            return scrobble

        scrobble.stop(force_finish=True)
        scrobble = cls.create(scrobble_data)
        logger.info(
            f"[scrobbling] finished - created for location",
            extra={
                "scrobble_id": scrobble.id,
                "media_id": location.id,
                "scrobble_data": scrobble_data,
                "media_type": cls.MediaType.GEO_LOCATION,
                "source": scrobble_data.get("source"),
            },
        )
        return scrobble

    def update(self, scrobble_data: dict) -> "Scrobble":
        # Status is a field we get from Mopidy, which refuses to poll us
        playback_status = scrobble_data.pop("playback_status", None)

        logger.info(
            "[scrobbling] update called",
            extra={
                "scrobble_id": self.id,
                "scrobble_data": scrobble_data,
                "media_type": self.media_type,
                "playback_status": playback_status,
            },
        )
        if self.beyond_completion_percent:
            playback_status = "stopped"

        if playback_status == "stopped":
            self.stop()
        if playback_status == "paused":
            self.pause()
        if playback_status == "resumed":
            self.resume()

        if playback_status != "resumed":
            scrobble_data["stop_timestamp"] = (
                scrobble_data.pop("timestamp", None) or timezone.now()
            )

        # timestamp should be more-or-less immutable
        scrobble_data.pop("timestamp", None)

        update_fields = []
        for key, value in scrobble_data.items():
            setattr(self, key, value)
            update_fields.append(key)
        self.save(update_fields=update_fields)

        logger.info(
            "[scrobbling] update finished",
            extra={
                "scrobble_id": self.id,
                "scrobble_data": scrobble_data,
                "playback_status": playback_status,
                "media_type": self.media_type,
            },
        )

        return self

    @classmethod
    def create(
        cls,
        scrobble_data: dict,
    ) -> "Scrobble":
        scrobble_data["log"] = {}
        scrobble = cls.objects.create(
            **scrobble_data,
        )
        return scrobble

    def stop(self, force_finish=False) -> None:
        self.stop_timestamp = timezone.now()
        self.played_to_completion = True
        self.in_progress = False

        if not self.playback_position_seconds:
            self.playback_position_seconds = int(
                (self.stop_timestamp - self.timestamp).total_seconds()
            )

        self.save(
            update_fields=[
                "in_progress",
                "played_to_completion",
                "stop_timestamp",
                "playback_position_seconds",
            ]
        )

        class_name = self.media_obj.__class__.__name__
        if class_name in LONG_PLAY_MEDIA.values():
            self.finish_long_play()

        logger.info(
            f"[scrobbling] stopped",
            extra={
                "scrobble_id": self.id,
                "media_id": self.media_obj.id,
                "media_type": self.media_type,
                "source": self.source,
            },
        )

    def pause(self) -> None:
        if self.is_paused:
            logger.warning(f"{self.id} - already paused - {self.source}")
            return
        self.is_paused = True
        self.save(update_fields=["is_paused"])
        logger.info(
            f"[scrobbling] paused",
            extra={
                "scrobble_id": self.id,
                "media_type": self.media_type,
                "source": self.source,
            },
        )

    def resume(self) -> None:
        if self.is_paused or not self.in_progress:
            self.is_paused = False
            self.in_progress = True
            self.save(update_fields=["is_paused", "in_progress"])
            logger.info(
                f"[scrobbling] resumed",
                extra={
                    "scrobble_id": self.id,
                    "media_type": self.media_type,
                    "source": self.source,
                },
            )

    def cancel(self) -> None:
        self.delete()

    def update_ticks(self, data) -> None:
        self.playback_position_seconds = data.get("playback_position_seconds")
        self.save(update_fields=["playback_position_seconds"])

    def finish_long_play(self):
        seconds_elapsed = (timezone.now() - self.timestamp).seconds
        past_seconds = 0

        # Set our playback seconds, and calc long play seconds
        self.playback_position_seconds = seconds_elapsed
        if self.previous:
            past_seconds = self.previous.long_play_seconds

        self.long_play_seconds = past_seconds + seconds_elapsed

        # Long play scrobbles are always finished when we say they are
        self.played_to_completion = True

        self.save(
            update_fields=[
                "playback_position_seconds",
                "played_to_completion",
                "long_play_seconds",
            ]
        )
        logger.info(
            f"[scrobbling] finishing long play",
            extra={
                "scrobble_id": self.id,
            },
        )

    @property
    def beyond_completion_percent(self) -> bool:
        """Returns true if our media is beyond our completion percent, unless
        our type is geolocation in which case we always return false
        """
        beyond_completion = (
            self.percent_played >= self.media_obj.COMPLETION_PERCENT
        )

        if self.media_type == "GeoLocation":
            logger.info(
                f"[scrobbling] locations are ONLY completed when new one is created",
                extra={
                    "scrobble_id": self.id,
                    "media_type": self.media_type,
                    "beyond_completion": beyond_completion,
                },
            )
            beyond_completion = False

        return beyond_completion